Abstract: For complete support of Quality of Service, it is better that environment itself predicts resource requirements of a job by using special methods in the Grid computing. The exact and correct prediction causes exact matching of required resources with available resources. After the execution of each job, the used resources will be saved in the active database named "History". At first some of the attributes will be exploit from the main job and according to a defined similarity algorithm the most similar executed job will be exploited from "History" using statistic terms such as linear regression or average, resource requirements will be predicted. The new idea in this research is based on active database and centralized history maintenance. Implementation and testing of the proposed architecture results in accuracy percentage of 96.68% to predict CPU usage of jobs and 91.29% of memory usage and 89.80% of the band width usage.
Abstract: Recently, there have been considerable efforts towards the convergence between P2P and Grid computing in order to reach a solution that takes the best of both worlds by exploiting the advantages that each offers. Augmenting the peer-to-peer model to the services of the Grid promises to eliminate bottlenecks and ensure greater scalability, availability, and fault-tolerance. The Grid Information Service (GIS) directly influences quality of service for grid platforms. Most of the proposed solutions for decentralizing the GIS are based on completely flat overlays. The main contributions for this paper are: the investigation of a novel resource discovery framework for Grid implementations based on a hierarchy of structured peer-to-peer overlay networks, and introducing a discovery algorithm utilizing the proposed framework. Validation of the framework-s performance is done via simulation. Experimental results show that the proposed organization has the advantage of being scalable while providing fault-isolation, effective bandwidth utilization, and hierarchical access control. In addition, it will lead to a reliable, guaranteed sub-linear search which returns results within a bounded interval of time and with a smaller amount of generated traffic within each domain.
Abstract: On a such wide-area environment as a Grid, data
placement is an important aspect of distributed database systems. In
this paper, we address the problem of initial placement of database
no-replicated fragments in Grid architecture. We propose a graph
based approach that considers resource restrictions. The goal is to
optimize the use of computing, storage and communication
resources. The proposed approach is developed in two phases: in the
first phase, we perform fragment grouping using knowledge about
fragments dependency and, in the second phase, we determine an
efficient placement of the fragment groups on the Grid. We also
show, via experimental analysis that our approach gives solutions
that are close to being optimal for different databases and Grid
configurations.
Abstract: A major requirement for Grid application developers is ensuring performance and scalability of their applications. Predicting the performance of an application demands understanding its specific features. This paper discusses performance modeling and prediction of multi-agent based simulation (MABS) applications on the Grid. An experiment conducted using a synthetic MABS workload explains the key features to be included in the performance model. The results obtained from the experiment show that the prediction model developed for the synthetic workload can be used as a guideline to understand to estimate the performance characteristics of real world simulation applications.
Abstract: The purpose of Grid computing is to utilize
computational power of idle resources which are distributed in
different areas. Given the grid dynamism and its decentralize
resources, there is a need for an efficient scheduler for scheduling
applications. Since task scheduling includes in the NP-hard problems
various researches have focused on invented algorithms especially
the genetic ones. But since genetic is an inherent algorithm which
searches the problem space globally and does not have the efficiency
required for local searching, therefore, its combination with local
searching algorithms can compensate for this shortcomings. The aim
of this paper is to combine the genetic algorithm and GELS (GAGELS)
as a method to solve scheduling problem by which
simultaneously pay attention to two factors of time and number of
missed tasks. Results show that the proposed algorithm can decrease
makespan while minimizing the number of missed tasks compared
with the traditional methods.
Abstract: The problem addressed herein is the efficient management of the Grid/Cluster intense computation involved, when the preconditioned Bi-CGSTAB Krylov method is employed for the iterative solution of the large and sparse linear system arising from the discretization of the Modified Helmholtz-Dirichlet problem by the Hermite Collocation method. Taking advantage of the Collocation ma-trix's red-black ordered structure we organize efficiently the whole computation and map it on a pipeline architecture with master-slave communication. Implementation, through MPI programming tools, is realized on a SUN V240 cluster, inter-connected through a 100Mbps and 1Gbps ethernet network,and its performance is presented by speedup measurements included.
Abstract: Grid computing is growing rapidly in the distributed
heterogeneous systems for utilizing and sharing large-scale resources
to solve complex scientific problems. Scheduling is the most recent
topic used to achieve high performance in grid environments. It aims
to find a suitable allocation of resources for each job. A typical
problem which arises during this task is the decision of scheduling. It
is about an effective utilization of processor to minimize tardiness
time of a job, when it is being scheduled. This paper, therefore,
addresses the problem by developing a general framework of grid
scheduling using dynamic information and an ant colony
optimization algorithm to improve the decision of scheduling. The
performance of various dispatching rules such as First Come First
Served (FCFS), Earliest Due Date (EDD), Earliest Release Date
(ERD), and an Ant Colony Optimization (ACO) are compared.
Moreover, the benefit of using an Ant Colony Optimization for
performance improvement of the grid Scheduling is also discussed. It
is found that the scheduling system using an Ant Colony
Optimization algorithm can efficiently and effectively allocate jobs
to proper resources.
Abstract: Workload and resource management are two essential functions provided at the service level of the grid software infrastructure. To improve the global throughput of these software environments, workloads have to be evenly scheduled among the available resources. To realize this goal several load balancing strategies and algorithms have been proposed. Most strategies were developed in mind, assuming homogeneous set of sites linked with homogeneous and fast networks. However for computational grids we must address main new issues, namely: heterogeneity, scalability and adaptability. In this paper, we propose a layered algorithm which achieve dynamic load balancing in grid computing. Based on a tree model, our algorithm presents the following main features: (i) it is layered; (ii) it supports heterogeneity and scalability; and, (iii) it is totally independent from any physical architecture of a grid.
Abstract: Trust management and Reputation models are
becoming integral part of Internet based applications such as CSCW,
E-commerce and Grid Computing. Also the trust dimension is a
significant social structure and key to social relations within a
collaborative community. Collaborative Decision Making (CDM) is
a difficult task in the context of distributed environment (information
across different geographical locations) and multidisciplinary
decisions are involved such as Virtual Organization (VO). To aid
team decision making in VO, Decision Support System and social
network analysis approaches are integrated. In such situations social
learning helps an organization in terms of relationship, team
formation, partner selection etc. In this paper we focus on trust
learning. Trust learning is an important activity in terms of
information exchange, negotiation, collaboration and trust
assessment for cooperation among virtual team members. In this
paper we have proposed a reinforcement learning which enhances the
trust decision making capability of interacting agents during
collaboration in problem solving activity. Trust computational model
with learning that we present is adapted for best alternate selection of
new project in the organization. We verify our model in a multi-agent
simulation where the agents in the community learn to identify
trustworthy members, inconsistent behavior and conflicting behavior
of agents.
Abstract: Distance visualization of large datasets often takes the direction of remote viewing and zooming techniques of stored static images. However, the continuous increase in the size of datasets and visualization operation causes insufficient performance with traditional desktop computers. Additionally, the visualization techniques such as Isosurface depend on the available resources of the running machine and the size of datasets. Moreover, the continuous demand for powerful computing powers and continuous increase in the size of datasets results an urgent need for a grid computing infrastructure. However, some issues arise in current grid such as resources availability at the client machines which are not sufficient enough to process large datasets. On top of that, different output devices and different network bandwidth between the visualization pipeline components often result output suitable for one machine and not suitable for another. In this paper we investigate how the grid services could be used to support remote visualization of large datasets and to break the constraint of physical co-location of the resources by applying the grid computing technologies. We show our grid enabled architecture to visualize large medical datasets (circa 5 million polygons) for remote interactive visualization on modest resources clients.
Abstract: Grid computing is a high performance computing
environment to solve larger scale computational applications. Grid
computing contains resource management, job scheduling, security
problems, information management and so on. Job scheduling is a
fundamental and important issue in achieving high performance in
grid computing systems. However, it is a big challenge to design an
efficient scheduler and its implementation. In Grid Computing, there
is a need of further improvement in Job Scheduling algorithm to
schedule the light-weight or small jobs into a coarse-grained or
group of jobs, which will reduce the communication time,
processing time and enhance resource utilization. This Grouping
strategy considers the processing power, memory-size and
bandwidth requirements of each job to realize the real grid system.
The experimental results demonstrate that the proposed scheduling
algorithm efficiently reduces the processing time of jobs in
comparison to others.
Abstract: Mobile devices, which are progressively surrounded
in our everyday life, have created a new paradigm where they
interconnect, interact and collaborate with each other. This network
can be used for flexible and secure coordinated sharing. On the other
hand Grid computing provides dependable, consistent, pervasive, and
inexpensive access to high-end computational capabilities. In this
paper, efforts are made to map the concepts of Grid on Ad-Hoc
networks because both exhibit similar kind of characteristics like
Scalability, Dynamism and Heterogeneity. In this context we
propose “Mobile Ad-Hoc Services Grid – MASGRID".
Abstract: Market based models are frequently used in the resource
allocation on the computational grid. However, as the size of
the grid grows, it becomes difficult for the customer to negotiate
directly with all the providers. Middle agents are introduced to
mediate between the providers and customers and facilitate the
resource allocation process. The most frequently deployed middle
agents are the matchmakers and the brokers. The matchmaking agent
finds possible candidate providers who can satisfy the requirements
of the consumers, after which the customer directly negotiates with
the candidates. The broker agents are mediating the negotiation with
the providers in real time.
In this paper we present a new type of middle agent, the marketmaker.
Its operation is based on two parallel operations - through
the investment process the marketmaker is acquiring resources and
resource reservations in large quantities, while through the resale process
it sells them to the customers. The operation of the marketmaker
is based on the fact that through its global view of the grid it can
perform a more efficient resource allocation than the one possible in
one-to-one negotiations between the customers and providers.
We present the operation and algorithms governing the operation
of the marketmaker agent, contrasting it with the matchmaker and
broker agents. Through a series of simulations in the task oriented
domain we compare the operation of the three agents types. We find
that the use of marketmaker agent leads to a better performance in the
allocation of large tasks and a significant reduction of the messaging
overhead.
Abstract: This paper is concerned with the design and implementation of MICOSim, an event-driven simulator written in Java for evaluating the performance of Grid entities (users, brokers and resources) under different scenarios such as varying the numbers of users, resources and brokers and varying their specifications and employed strategies.
Abstract: This paper is to present context-aware sensor grid
framework for agriculture and its design challenges. Use of sensor
networks in the domain of agriculture is not new. However, due to
the unavailability of any common framework, solutions that are
developed in this domain are location, environment and problem
dependent. Keeping the need of common framework for agriculture,
Context-Aware Sensor Grid Framework is proposed. It will be
helpful in developing solutions for majority of the problems related
to irrigation, pesticides spray, use of fertilizers, regular monitoring of
plot and yield etc. due to the capability of adjusting according to
location and environment. The proposed framework is composed of
three layer architecture including context-aware application layer,
grid middleware layer and sensor network layer.
Abstract: A Web-services based grid infrastructure is evolving to be readily available in the near future. In this approach, the Web services are inherited (encapsulated or functioned) into the same existing Grid services class. In practice there is not much difference between the existing Web and grid infrastructure. Grid services emerged as stateful web services. In this paper, we present the key components of web-services based grid and also how the resource discovery is performed on web-services based grid considering resource discovery, as a critical service, to be provided by any type of grid.
Abstract: Some meta-schedulers query the information system of individual supercomputers in order to submit jobs to the least busy supercomputer on a computational Grid. However, this information can become outdated by the time a job starts due to changes in scheduling priorities. The MSR scheme is based on Multiple Simultaneous Requests and can take advantage of opportunities resulting from these priorities changes. This paper presents the SWARM meta-scheduler, which can speed up the execution of large sets of tasks by minimizing the job queuing time through the submission of multiple requests. Performance tests have shown that this new meta-scheduler is faster than an implementation of the MSR scheme and the gLite meta-scheduler. SWARM has been used through the GridQTL project beta-testing portal during the past year. Statistics are provided for this usage and demonstrate its capacity to achieve reliably a substantial reduction of the execution time in production conditions.